Vision And TOF Based Navigation For UAV

In the last decade,civil unmanned aerial vehicle(UAV)has changed a great deal.With the number of UAV increased,sense and avoid has been increased interest.Therefore,autopilot system,with robust and accurate,is a necessary model of UAV.With the development of(micro electromechanical systems(MEMS),UAV's autopilot system has become cheap and efficient.However,we have to solve several problems to achieve greater security.For example,sense and avoid,work without GPS,etc.Most of the micro UAVs are based on GPS and Inertial Navigation System(INS),don't have the ability to sense and avoid.These systems can't work in many cases(indoor,etc.),Now the number of the techniques have been designed for these cases.In these paper,we conclude the main approaches and present two navigation system with the vision and rangefinder.The vision system consists of the vision model and the flight controller.System get the real-time velocity from the vision model and remain in a steady position.The rangefinder system includes TOF rangefinders with different direction on UAV.System could sense the environment around the aircraft.And then,the flight controller could navigate the vehicle autonomously.After the test,we known that the vision system could keep the UAV's position without the compass and GPS.The TOF rangefinders system could control the UAV's position and yaw angle autonomously.And the next task,we would use the more complex vision calculation to get better results.And we could improve the structure of TOF hardware to get more better performance.After that,the vision and the TOF sensors could work with other sensors,like millimeter-wave radars and arrays TOF.